EARLY REPORT

Early report

Coronary angiography with multi-slice computed tomography Koen Nieman, Matthijs Oudkerk, Benno J Rensing, Peter van Ooijen, Arie Munne, Robert-Jan van Geuns, Pim J de Feyter

Summary Background A new generation of subsecond multi-slice computed tomography (MSCT) scanners, which allow complete coronary coverage, are becoming widely available. We investigated the potential value of MSCT angiography in a range of coronary disorders. Methods We studied 35 patients, including 11 who had undergone percutaneous transluminal coronary angioplasty and four who had had coronary-artery bypass grafts, by both MSCT and conventional coronary angiography. After intravenous injection of a non-ionic contrast medium with high iodine content, the entire heart was scanned within a single breath-hold. The total examination time was no more than 20 min. The retrospective electrocardiographically gated reconstruction source images and three-dimensional reconstructed volumes were analysed by two investigators, unaware of the results of conventional angiography. Findings In the 31 patients without previous coronary surgery, 173 (73%) of the 237 proximal and middle coronary segments were assessable. In the assessable segments, 17 of 21 significant stenoses (>50% reduction of vessel diameter) were correctly diagnosed. The non-assessable segments included four lesions. Misinterpretations were mainly the result of severe calcification of the vessel wall. Segments with implanted stents were poorly visualised, but stent patency could be assessed in all cases. Of the 17 segments of bypass grafts, 15 were assessable and four of five graft lesions were detected. Two cases of anomalous coronary anatomy could be visualised well. Interpretation These preliminary data suggest that MSCT allows non-invasive imaging of coronary-artery stenoses and has potential to develop into a reliable clinical technique. Lancet 2001; 357: 599–603

Introduction Conventional X-ray coronary angiography is the standard of reference for the assessment of coronary-artery disease. It is an invasive and potentially harmful procedure with a small risk of serious events (arrhythmia, stroke, coronaryartery dissection, death). Furthermore, the catheterisation procedure involves admission to hospital and discomfort for the patient. Therefore, conventional angiography should be undertaken only on strict clinical indications. Magnetic resonance imaging and electron-beam computed tomography have been investigated for noninvasive coronary imaging. However, both have significant Department of Cardiology, Thoraxcenter (K Nieman MD, B J Rensing MD, R-J van Geuns MD, P J de Feyter MD) and Department of Radiology, Daniel Den Hoed Kliniek (M Oudkerk MD, P van Ooijen MSc, A Munne RT), Rotterdam University Hospital, Rotterdam, Netherlands Correspondence to: Dr Pim J de Feyter, Catheterisation Laboratory, Thoraxcenter BD 406, PO Box 2040, 3000 CA Rotterdam, Netherlands (e-mail: [email protected])

THE LANCET • Vol 357 • February 24, 2001

limitations in reliable visualisation of the coronary arteries.1–4 Multi-slice computed tomography (MSCT) scanners have lately become widely available. These scanners have the potential to allow non-invasive coronary angiography within a single breath-hold by use of a rotation speed of 0·5 s and sophisticated algorithms for retrospective electrocardiographic (ECG) gating. We aimed in this study to assess the diagnostic potential of non-invasive MSCT angiography for the assessment of coronary-artery disease.

Methods 35 patients (27 male, eight female; mean age 59 years [SD 11; range 28–77]) underwent both conventional and MSCT angiography of the coronary arteries or bypass grafts. 11 patients had previously undergone percutaneous transluminal coronary angioplasty with stent implantation, and four coronary-artery bypass grafting. Two patients had a congenital coronary-artery variant. Patients were included only if they had a regular heart rhythm and good pulmonary function. Exclusion criteria were: previous allergic reactions to iodine-containing contrast media, severe renal failure, pregnancy, an unstable clinical condition, or circumstances of any kind that would not allow the patient to lie in a supine position. All patients scheduled for conventional coronary angiography at our centre were approached until the scanning slots allocated to the research project were filled. The study protocol was approved by the hospital’s ethics committee and all patients gave informed consent. A new generation of multi-detector-array CT scanners operate at an increased rotation rate (2 per s) and produce up to four slices simultaneously.5 These developments permit high-speed scanning of large volumes with a high in-plane resolution, as well as an improved Z-axis resolution and a substantial improvement in the inter-slice correlation.6 Partial scan-reconstruction techniques, which apply a 90–180º reconstruction algorithm, improve the temporal resolution to 250 ms. These algorithms also reduce the effective slice thickness at an acceptable increase in noise and reduction in contrast.7 Recent modifications in the reconstruction software have further reduced the virtual temporal resolution at higher heart rates by combining the data from several heart cycles in one image, shortening the effective acquisition intervals to 125 ms. The in-plane spatial resolution of the MSCT scanner is nine line pairs per cm.8 We used retrospective gating, which allows post-scan acquisition window selection and optimum gating.9 This approach improves the image quality and decreases the sensitivity to arrhythmia and ECG noise. The patient was placed within the gantry of an MSCT scanner (Somatom plus 4 Volume Zoom, Siemens AG, Erlangen, Germany) in a supine position. Leads were attached for simultaneous ECG and image recording necessary for inter-related image reconstruction. According to the expected location of the coronary arteries, obtained from the coronal scout view, the scan volume was defined, depending on the patient’s ability to cooperate and characteristics (breath-hold, heart rate) and

599

For personal use only. Reproduce with permission from The Lancet Publishing Group.

EARLY REPORT

the scan variables (pitch, slice thickness, scan time). For arterial grafts, the area to be covered was extended to the origin of the left internal mammary artery. Fixed scanning variables included the 0·5 s rotation time and a tube voltage of 140 kV. We used a protocol of four slices with a collimated slice thickness of 1 mm. The pitch (table feed per rotation divided by the single collimated slice thickness) was set at 1·5 for heart rates below 80 beats per min and 2·0 for faster heart rates. Depending on the selected pitch and volume to be covered, the total scan time was between 30 s and 40 s. After thorough instruction, most of the patients were able to suspend respiration for 40 s. To facilitate adequate breath-holding, the patients were asked to hyperventilate before the start of the scan. Optimum contrast between blood and surrounding tissue was achieved by injection of 150 mL contrast agent with a high iodine content (350 g iodine per L; Iomeprol, Bracco-Byk Gulden, Konstanz, Germany) into the antecubital vein at a rate of 3·5–4·0 mL/s. Scanning was started after 20 s. The radiation dose was estimated to be 4·9 mSv (ICRP 60, Monte Carlo, Aarhuys University Hospital). The acquired CT and ECG data were sent to a separate workstation, and dedicated cardiac work-in-progress reconstruction software (Siemens Cardio Package, Siemens AG; and MatLab version 5.3, MathWorks Inc, Natick, MA, USA) was used to reconstruct the images. Transverse tomograms were reconstructed from the acquired CT data during a preselected interval of 125–250 ms, depending on the heart rate, within the cardiac cycle. To minimise motion artefacts, the reconstruction window was positioned in the mid to late diastolic phase at a fixed point before the next R wave. If ECG irregularities occur, the retrospective ECG gating can be corrected manually. Owing to the spiral motion of the detector row, the 41 mm collimation results in an effective slice thickness of 1·25 mm. Through reconstruction of overlapping slices at an increment of 0·8 mm, a stack of 130–150 slices is created. Further parameters include a 150 mm field of view and a 512512 matrix, which results in an interpolated voxel size of 0·30·30·8 mm3.

The images were further processed on separate graphic workstations (Indigo 2 and O2, SGI, Mountain View, CA) by means of special software packages (Vitrea and Voxel View, Vital Images, Plymouth, MN, USA). To analyse the coronary arteries, several volume-rendering techniques were used. Multiplanar reformatting allows the investigator to place and manoeuvre cross-sectional image planes through a three-dimensional volume. High-level (>150 Hounsfield units) and narrow window settings were used to discriminate between contrastenhanced lumen and the vessel wall. Three-dimensional volume-rendering techniques with manual segmentation of overlying structures (Voxel View, Vital Images) were used to demonstrate the three-dimensional course of the cardiac vessels around the heart. All scans were evaluated by consensus of two experienced investigators, who were unaware of the conventional angiographic results. If consensus could not be reached, a third investigator was consulted. Cardiac catheterisation and contrast-enhanced X-ray coronary angiography were done according to standard techniques. Multiple views of the coronary arteries were obtained and stored on a CD-ROM. The angiograms were evaluated by two cardiologists without knowledge of the MSCT angiographic findings. In cases of disagreement, a third cardiologist was consulted. Coronary-artery segments were classified as significantly stenosed (diameter reduction >50%) or as normal or not significantly stenosed (diameter reduction 50%). We investigated the proximal and middle segments of the coronary-artery tree, which includes the proximal, middle, and distal segments of the right coronary artery, the left main artery, the proximal and middle segments of the left anterior descending artery, and the proximal and middle segments of the left circumflex artery, according to the guidelines of the American Heart Association.10 Thus, eight segments per patient were available for assessment. After localisation, the respective segments were first semiquantitatively classified as assessable or not. Segments with stents were excluded. Results from the two angiographic techniques were compared, with conventional angiography serving as the standard of reference.

Figure 1: Left anterior oblique projection with cranial angulation by MSCT coronary angiography (A) and conventional coronary angiography (B) Occlusion (black arrow) of the left anterior descending artery (LAD) with collateral filling of the distal LAD as well as a severe stenosis (white arrow) of the second diagonal branch (D2), were detected with both modalities. The septal branch (S1) and the first diagonal artery (D1) are well visualised. The proximal left circumflex artery (LCX) appears severely calcified on the MSCT angiogram. LM=left main coronary artery.

600

THE LANCET • Vol 357 • February 24, 2001

For personal use only. Reproduce with permission from The Lancet Publishing Group.

EARLY REPORT

Segment

Number assessable/total

Right coronary artery Proximal Middle Distal

26/29 (90%) 18/30 (60%) 20/31 (65%)

Left main artery

29/30 (97%)

Left anterior descending artery Proximal Middle

28/30 (93%) 20/26 (77%)

Left circumflex artery Proximal Middle Total

21/30 (70%) 11/31 (35%) 173/237 (73%)

11 segments with an implanted stent were excluded.

Table 1: Number of assessable segments in 31 patients

Results The median time between conventional and MSCT angiography was 9 days (range 0–39). During the MSCT investigations, no severe complications occurred. Two patients developed an allergic skin reaction, one within 1 h and the other 48 h after injection of the contrast agent.

Figure 2: Conventional and MSCT angiography of left anterior descending artery (LAD) A: Conventional X-ray coronary angiogram shows a stenosis (black arrow) in LAD proximal to the first diagonal artery (D1). B: Owing to extensive calcification, the stenosis is obscured (white arrow) on the MSCT angiogram. C: High-density threshold MSCT representation of the LAD, which exclusively shows the calcifications in the vessel wall. LM=left main coronary artery; LCX=left circumflex artery; S1=first septal branch; GCV=great cardiac vein.

THE LANCET • Vol 357 • February 24, 2001

Cardiac motion/arrhythmia Extensive calcifications Small vessel (50% diameter

601

For personal use only. Reproduce with permission from The Lancet Publishing Group.

EARLY REPORT

Two patients presented with anomalous anatomy of the coronary vessels. In both cases the coronary anatomy could be presented in a readily interpretable threedimensional image (figure 4). Video recordings of four angiograms are available with this paper on The Lancet’s website (www.thelancet.com).

Discussion We found that significant stenoses (>50% reduction in diameter) in the proximal and middle coronary arteries could be detected by MSCT coronary angiography. Segments of the left main and left anterior descending arteries were visualised in most cases (90%), and six of eight stenoses were correctly detected. Visualisation of the right coronary artery is more sensitive to motion artefacts, which resulted in a lower proportion of interpretable segments (71%). Nevertheless, all seven lesions were correctly diagnosed. The left circumflex artery was the most dificult to examine. This artery is small in many people, and it easily blends with adjacent contrast-filled structures such as the great cardiac vein and the left atrium. Only 51% of the circumflex segments were assessable, and two of four stenoses were detected. In the presence of intracoronary stents, high-density artefacts, combined with partial volume effects, prevent adequate assessment of the small vessel lumen within the struts of the stent. However, patency can be assessed if enhancement by contrast medium is observed in the vessel segment distal to the stent. In larger-diameter vessels, such as the carotid arteries, the in-stent lumen can be assessed quite well.11 Assessability of the stented

segments is expected to increase with the improvement of spatial resolution and the use of less radio-opaque alloys for stents. Coronary-artery bypass grafts, because of their size and relative immobility, can be reliably imaged. In patients with an anomalous origin or course of the coronary arteries, MSCT displays, in contrast to conventional angiography, a three-dimensional map of the coronary anatomy, which allows easy identification of a high-risk course of a coronary artery between the pulmonary artery and the aorta (figure 4). Despite these satisfying initial results, some technical limitations remain. Although manual repositioning of the R-wave indicators during retrograde gating improves the synchronisation of acquisition intervals between consecutive heartbeats, cardiac motion artefacts cannot be entirely prevented. For instance, the middle segment of the right coronary artery, which is mobile during the cardiac cycle, runs perpendicular to the transverse slices. Consequently, this vessel is more vulnerable to arrhythmia and inaccurate triggering, which results in discontinuity between the consecutive slices. Movement of the patient, such as breathing, also causes motion artefacts, which can be reduced by thorough instruction before scanning. The presence of extensive calcifications can complicate correct assessment of the lumen of the coronary arteries. The high-contrast calcium depositions cannot be sufficiently isolated from the contrast-enhanced vessel lumen and may result in non-assessable segments or misinterpretation. Nevertheless, severe calcification of the coronary arteries is related to coronary-artery disease, and its detection will contribute to clinical decision-making.12

Figure 3: Conventional coronary angiography (A, B, C) and MSCT angiography (D, E) of a patient with previous coronary-artery bypass grafting Left internal mammary artery (LIMA) is anastomosed to the left anterior descending artery (LAD). A saphenous vein graft (SVG) jumps via the marginal branch (RM) to the posterolateral branch (RPL) and the right descending posterior branch (RDP). The last segment shows a significant lesion (arrow head). SC=coronary sinus; GCV=great cardiac vein.

602

THE LANCET • Vol 357 • February 24, 2001

For personal use only. Reproduce with permission from The Lancet Publishing Group.

EARLY REPORT

spatial orientation of vessels and are able to identify and quantify calcium deposition within the vessel wall. Contributors Koen Nieman contributed to the study concept and design, literature research, clinical studies, and data acquisition and analysis, and was responsible by preparation of the report. Mathijs Oudkerk contributed to concept, design, analysis, and editing and review of the report. Benno Rensing contributed to concept, design, literature research, analysis, and editing and review of the report. Peter van Ooijen contributed to data acquisition and review of the report. Arie Munne contributed to design, clinical studies, data acquisition, and review of the report. Robert van Geuns contributed to concept, design, clinical studies, and review of the report. Pim de Feyter contributed to concept design, literature research, data analysis, and editing and review of the report.

References 1 2

3

4

Figure 4: Three-dimensional reconstruction (cranio-left-anterior view) of an MSCT scan of a coronary anomaly Aberrant right coronary artery (RCA) originates from left aortic sinus near left main coronary artery and runs between aorta (AO) and pulmonary artery, which has been removed, to right atrioventricular groove. This notorious anomaly can lead to myocardial infarction and sudden death. RVOT=right ventricle outflow tract; LA=left atrium; LAD=left anterior descending artery; D1=first diagonal artery; MO=marginal branch.

Magnetic resonance coronary angiography can visualise the coronary anatomy and detect stenotic lesions in the proximal segments of the coronary arteries.13–18 However, the diagnostic accuracy varies substantially between studies.1,19 The advantages of magnetic resonance angiography are the absence of ionising radiation and iodine contrast agents and the opportunity to combine the assessment of the coronary arteries with the examination of other features related to ischaemic heart disease, such as ventricular function, myocardial perfusion, and coronary flow.20 However, this technique is still hampered by poor spatial resolution, long scan times, image degeneration by metal objects (stents, sternal wires), and contraindications to magnetic resonance imaging. The electron-beam CT scanner is a non-mechanical sequential CT scanner, in which the electron beam that produces the X-rays rotates around the patient. Because there are no mechanically rotating components, the temporal resolution (100 ms) is high. However, the inplane spatial resolution is slightly lower than in MSCT.8 Electron-beam CT is prospectively gated. When a 1·5 mm slice thickness protocol is applied, it can cover 6 cm in 40 heart beats, which allows only scanning of the proximal and middle segments of the coronary arteries of a normal-sized heart.2,3 The entire heart can be covered with a 3 mm protocol at the cost of a reduction in Z-axis resolution. The non-invasive modalities still fall short of the diagnostic superiority of conventional coronary angiography, but these new techniques are only in the early stages of development. Further technical refinement of the individual modalities and introduction of computer systems that allow significantly faster data-processing will accelerate the clinical implementation of non-invasive coronary angiography. Besides being a non-invasive alternative, MSCT and electron-beam CT offer additional information about the

THE LANCET • Vol 357 • February 24, 2001

5

6

7

8

9

10 11

12

13

14

15

16

17

18

19

20

Wielopolski PA, van Geuns RJ, de Feyter PJ, Oudkerk M. Coronary arteries. Eur Radiol 2000; 10: 12–35. Rensing BJ, Bongaerts A, van Geuns RJ, et al. Intravenous coronary angiography by electron beam computed tomography: a clinical evaluation. Circulation 1998; 98: 2509–12. Achenbach S, Moshage W, Ropers D, Nossen J, Daniel WG. Value of electron-beam computed tomography for the noninvasive detection of high-grade coronary-artery stenoses and occlusions. N Engl J Med 1998; 339: 1964–71. de Feyter PJ, Nieman K, van Ooijen P, Oudkerk M. Non-invasive coronary artery imaging with electron beam computed tomography and magnetic resonance imaging. Heart 2000; 84: 442–48. Klingenbeck-Regn K, Schaller S, Flohr T, Ohnesorge B, Kopp AF, Baum U. Subsecond multi-slice computed tomography: basics and applications. Eur J Radiol 1999; 31: 110–24. Hu H, He HD, Foley WD, Fox SH. Four multidetector-row helical CT: image quality and volume coverage speed. Radiology 2000; 215: 55–62. Kachelreiss M, Kalender WA. Electrocardiogram-correlated image reconstruction from subsecond spiral computed tomography scans of the heart. Med Phys 1998; 25: 2417–31. Becker CR, Knez A, Leber A, et al. Erste Erfahrungen mit der Mehrzeilendetektorspiral-CT in der Diagnostik der Arteriosklerose der Koronargefasse. Radiologe 2000; 40: 118–22. Ohnesorge B, Flohr T, Becker C, et al. Cardiac imaging by means of electrocardiographically gated multisection spiral CT: initial experience. Radiology 2000; 217: 564–71. Austen WG, Edwards JE, Frye RL, et al. A reporting system on patients evaluated for coronary artery disease. Circulation 1975; 51: 5–40. Leclerc X, Gauvrit JY, Pruvo JP. Usefulness of CT angiography with volume rendering after carotid angioplasty and stenting. AJR Am J Roentgenol 2000; 174: 820–22. Bielak LF, Rumberger JA, Sheedy PF II, Schwartz RS, Peyser PA. Probabilistic model for prediction of angiographically defined obstructive coronary artery disease using electron beam computed tomography calcium score strata. Circulation 2000; 102: 380–85. Post JC, van Rossum AC, Hofman MB, Valk J, Visser CA. Three-dimensional respiratory-gated MR angiography of coronary arteries: comparison with conventional coronary angiography. AJR Am J Roentgenol 1996; 166: 1399–404. Pennell DJ, Bogren HG, Keegan J, Firmin DN, Underwood SR. Assessment of coronary artery stenosis by magnetic resonance imaging. Heart 1996; 75: 127–33. Huber A, Nikolaou K, Gonschior P, Knez A, Stehling M, Reiser M. Navigator echo-based respiratory gating for three-dimensional MR coronary angiography: results from healthy volunteers and patients with proximal coronary artery stenoses. AJR Am J Roentgenol 1999; 173: 95–101. Sandstede JJ, Pabst T, Beer M, Geis N, Kenn W, Neubauer S, Hahn D. Three-dimensional MR coronary angiography using the navigator technique compared with conventional coronary angiography. AJR Am J Roentgenol 1999; 172: 135–39. Kessler W, Achenbach S, Moshage W, et al. Usefulness of respiratory gated magnetic resonance coronary angiography in assessing narrowings > or=50% in diameter in native coronary arteries and in aortocoronary bypass conduits. Am J Cardiol 1997; 80: 989–93. van Geuns RJM, de Bruin HG, Wielopolski PA, et al. MRI of the coronary arteries: clinical results from three dimensional evaluation of a respiratory technique. Heart 1999; 82: 515–19. van Geuns RJM, Wielopolski PA, de Bruin HG, et al. MR coronary angiography with breath-hold targeted volumes: preliminary clinical results. Radiology 2000; 217: 270–77. Pattynama PM, De Roos A, Van der Wall EE, Van Voorthuisen AE. Evaluation of cardiac function with magnetic resonance imaging. Am Heart J 1994; 128: 595–607.

603

For personal use only. Reproduce with permission from The Lancet Publishing Group.